Viral gene sequences represent a rich and valuable source of information about biological processes. Advances in next generation sequencing (NGS) technology over the past decade now provide the opportunity to probe viral population diversity and evolution in unprecedented ways. We developed specialized sequencing strategies to overcome several serious limitations of conventional NGS in analyzing viral populations, including greatly reducing the mis- incorporation and recombination introduced by the preceding PCR step, reducing the errors of the sequencing platform, and revealing sampling depth of the initial viral genomes/templates that are actually represented in the final data set. In this application we will use state-of-the-art NGS to address long-standing issues in HIV-1 population analysis in the context of viral evolution in the absence of therapy, the role of minor variants in predicting failed therapy, and whether the latent reservoir on therapy is replicating. In addition, we will link our sequence data to state-of-the-art evolutionary analysis, and confirm key aspects of our inferences with measures of changing viral phenotypes and host environment. In Aim 1, we will analyze longitudinal plasma samples from 27 HIV-infected women (from the WIHS cohort) starting with high CD4+ T cell counts until they progress to CD4+ T cell counts of less than 100 cells/L. We will perform multiplexed NGS sequencing to obtain near full length HIV-1 genome sequences. We predict that X4 variants in viral populations first emerge at low abundance and that we will be able to detect them much earlier than previously observed and follow their evolution. In addition, we will investigate longitudinal viral diversity changes in all sequenced regions to assess population dynamics and link these changes to markers of inflammation, CNS damage, CTL response, and the breadth and potency of neutralizing antibodies. In Aim 2, we will apply multiplexed NGS sequencing as a screening tool for minor drug resistance variants that predict therapy failure. We hypothesize that the potential for drug resistance mutations to mediate escape can occur from minor variants, too minor to be reliably detected with the methods that have been used to date. We will analyze samples from four cohorts (Malawi and China) to determine how often minor variants are missed by using Sanger sequencing. We will then link resistance mutations leading to therapy failure with their pretherapy abundance in a case- control design. In Aim 3, we will use near full length genome sequencing of reservoir virus (either outgrowth or rebound) in 13 participants who were infected with a single variant and started therapy early. Potential evolution on therapy will include a focus on extant CTL responses. In this way we will provide a critical test of sequence evolution as a signature of viral replication during years of successful suppressive therapy. With this application we are examining questions that are central to understanding HIV-1 in the absence of therapy, when faced with the selection of antiretroviral drugs, and on successful therapy. Insights into each of these questions can be first approached with the appropriate use of NGS sequencing but in ways that account for the strengths and weaknesses of these platforms. High quality sequence information then leads the way to insights in viral evolution in response to a changing host, in response to drug selection, and in response to apparently suppressive therapy.